Primary osteosarcomas of the mandible: Joint study of the Groupe Sarcome Francais (GSF), Rare Cancer Network, and Groupe d’Etude des Tumeurs de la Tete et du Cou (GETTEC).

Abstract

10036 Background: Mandibular osteosarcoma (MOS) is a rare disease (dz) whose incidence is 1/10 million/year. The prognosis mainly depends on local control. Indications of radiotherapy (RT) for R1 disease (dz) or chemotherapy (chemo) vary among centers. This study focused on the impact of surgery and perioperative management toward outcome. Methods: Prognostic factors (age, histology, tumor size, margins, RT and chemo) for local control and survival were analyzed. Results: 93 patients (pts) treated between 1994 and 2004 were included; 63% were male. Median age was 37 years (14-79). Median tumor size at diagnosis was 4.5 cm. Nodal involvement occurred in 4%. Histology was conventional 32%, osteoblastic 23%, chondroblastic 20%, fibroblastic 5%, small cell 1%, low-grade 4% or unspecified 15%. Tumor size influenced the use of preoperative chemo (p .017) and there was a center effect for chemo and RT use (p .015). Preoperative chemo was performed in 71%, with various regimens containing adriamycin, cisplatin, and methotrexate, yielding a 30% favorable histological response rate. Adequate margins (≥ 0.5 cm) were obtained in 56% of pts. Preoperative chemo improved the rate of R0 margins (68% vs 37%) (p .045). Postoperative chemo was delivered in 58% and postoperative RT in 26% of pts: 42% / 58% with R0 / R1-2 dz, respectively. Neither in R0 nor R1 groups did RT (median 60 Gy) improve local control. Median follow-up for living pts was 52 months. Crude local control, metastatic failure, dz-free, survival rates were 71%, 26%, 63% and 65%, respectively. Metastatic rates were 24% / 32% with / without chemo (NS) and 44% / 19% with / without local relapse (p .01). On uni/multivariate analyses, margins (2 cm / 0.5-2 cm / R1 / R2) were prognostic factors for local control and dz-free survival (p .015); age and tumor size for overall survival (p .05). Conclusions: This largest series to date mostly included high grade MOS. Despite being non-randomized, the results suggest that preoperative chemo may increase the rate of adequate margins, which was a major prognostic factor for local control and dz-free survival. The subset of pts that might benefit from RT and the optimal dose remain to be identified.